Assessing the impact of knowledge-based planning on dosimetric performance in adaptive prostate stereotactic body...

Ethos Intelligent Optimization Engine version 2 (IOE v2) generates automated treatment plans using user-defined templates that may not fully account for patient-specific anatomy. This retrospective study evaluated whether incorporating RapidPlan knowledge-based planning guidance into the Ethos IOE v2 workflow could improve dosimetric quality for adaptive prostate stereotactic body radiation therapy (SBRT). Fifteen patients previously treated with adaptive prostate SBRT were retrospectively evaluated. Three plans were generated for each patient: a reference Ethos IOE v2 plan without RapidPlan guidance (Ref), a plan using an initial RapidPlan model trained on 20 cases (RP1), and a plan using a refined model trained on 35 plans, including the 15 evaluation cases, and reapplied to those same cases (RP2). Plans were compared using dose-volume histogram metrics for target coverage and organ-at-risk sparing, as well as total monitor units (MU). Paired t tests or 2-sided sign tests were used based on the normality of paired differences, with statistical significance defined as p < 0.05. Compared with Ref, RP1 and RP2 reduced bladder mean dose by 0.74 Gy and 0.86 Gy, respectively, and reduced bladder V15 Gy by 2.80 and 3.35 percentage points. Both RapidPlan-guided approaches also significantly reduced penile bulb and sigmoid dose metrics. Urethral maximum dose increased by 0.44 Gy with RP1 and 0.35 Gy with RP2 but remained within the institutional clinical goal of 40 Gy for all cases. Small but statistically significant reductions in PTV_3625 D95% and D98% were observed, whereas PTV_mb_4250 coverage was maintained. Mean MU decreased from 3415.1 for Ref to 2816.9 for RP1 (p < 0.001) and 2836.3 for RP2 (p = 0.007). RapidPlan guidance improved selected organ-at-risk dose metrics and reduced MU in Ethos IOE v2 adaptive prostate SBRT planning. However, these improvements were accompanied by small reductions in PTV_3625 coverage and increases in urethral maximum dose. Further model refinement and independent validation are needed to better balance target coverage and organ-at-risk sparing.